Genomic Heterogeneity in ctDNA During Therapy Affects Prognosis of mCRC

Molecular screening for actionable alterations, such as RAS, BRAFV600E, and MMR/MSI in metastatic colorectal cancer (mCRC), is key for personalized treatment selection that improves patient outcomes. Advances in technologies, such as next-generation sequencing (NGS), have shown that CRC is a highly heterogeneous disease. Furthermore, it has been demonstrated that molecular heterogeneity, which develops over time (temporal heterogeneity), can cause treatment resistance. Although tissue biopsy is the standard of care for molecular testing, plasma circulating tumor DNA (ctDNA) sequencing (liquid biopsy) is a non-invasive alternative, which is becoming increasingly used throughout the course of CRC management. It is well known that ctDNA can track RAS clones in mCRC and may guide treatment selection with anti-EGFR rechallenge. However, it is largely unknown how ctDNA is capturing the evolution of somatic mutations during systemic therapy and what the impact is of this genomic heterogeneity on patient outcomes.

A group of researchers from the Sun Yat-Sen University Cancer Center (Guangzhou, China) sought to investigate these questions in a prospective cohort study of 171 patients with unresectable mCRC who underwent first-line systemic therapy. Baseline RAS mutations were detected in ctDNA in 74 (43.3%) patients, and BRAF mutations were detected in 11 (6.4%) patients. Among 63 patients with paired baseline plasma and tissue samples, the overall agreement of RAS/BRAFV600E status between ctDNA and tissue was 81%, with 17.5% disagreement in RAS status and 1.5% disagreement in BRAF status. Interestingly, concordance rates varied between metastatic sites (90% in patients with liver-only metastases, and 79.1% in those with extrahepatic metastases). Genomic evolution under first-line treatment was assessed in a cohort of 145 patients with sequentially collected plasma samples before progressive disease (PD). After a period of first-line treatment (median 4.7 months), there was a decline in mutation frequencies in the majority of genes, with the most dramatic shift being observed in patients with RAS and BRAF mutational status. RAS mutation clearance was found in 42.6% of patients with RAS mutations and BRAF mutation clearance in 50% of patients with BRAF mutations, while 3.6% and 0.7% of patients gained new RAS and BRAF mutations, respectively.

Importantly, these genomic changes in RAS and BRAF were correlated with patient outcomes. The median progression-free survival (mPFS) in patients with RAS or BRAF clearance (12.8 months and not reached [NR]) were similar to mPFS of patients who remained RAS or BRAF wild-type (13.2 and 11.5 months, respectively) and were significantly better compared with patients who remained RAS or BRAF mutant (7.9 months and 6.4 months, respectively). Likewise, patients with RAS or BRAF clearance had similar median overall survival (mOS) to RAS and BRAF wild-type patients (NR vs. 31.4 and 26.7 months, respectively), but patients who remained RAS or BRAF mutant had a significantly shorter mOS (14.5 and 11.4 months). Patients who developed new RAS or BRAF mutations had a similar prognosis to patients who remained RAS or BRAF mutant and had a shorter PFS and OS than those who remained wild-type. The ctDNA levels decreased in patients with partial response or stable disease and increased in patients with disease progression. This indicates the value of ctDNA use in response evaluation. In addition, investigators observed significant changes in somatic mutations at the time of treatment failure on first-line therapy. This information could be helpful to determine subsequent therapies.

In their conclusion, the authors highlighted that this prospective study provides further evidence supporting the use of ctDNA sequencing to capture the heterogeneity and temporal genomic dynamics of mCRC. This is clinically relevant as the shift in plasma RAS and BRAF mutational status appears predictive for survival.

Wang F, et al. Gut. 2021; Sep 6 [Online ahead of print].